PSYB04H3 Lecture Notes - Lecture 11: Systematic Review, Positive Psychology, Internal Validity
Document Summary
Lecture 11 importance of studies (replicability, generalizability and realism) Statistical power: alpha ( ) = probability of type i error (blue area below) 1 = pro(cid:271)a(cid:271)ility of (cid:272)orre(cid:272)t de(cid:272)isio(cid:374) whe(cid:374) the (cid:374)ull is true eg. ) no effect exists in reality: beta (cid:894) (cid:895): pro(cid:271)a(cid:271)ility of a type ii error (cid:894)pi(cid:374)k area (cid:271)elow(cid:895) (1 (cid:895) = pro(cid:271)a(cid:271)ility of (cid:272)orre(cid:272)t de(cid:272)ision when the null is not true eg. ) there is an effect in real life. This is power: statistical power: probability of finding effect if one exists in reality, function of effect size, type i error rate, sample size and rm designs. When alpha is not firm: alpha conventionally set at . 05 or . 01. > adds weight to findings, helps identify fraud, overcomes problems with, small sample and specialized methods. Why is(cid:374)"t it do(cid:374)e (cid:373)ore: not rewarded (no glory or funding) and not published (null results) Mata-analysis: what does the literature say: synthesis of all research findings in an area.
